Dear LF Group,
To try to get an idea of the order of magnitude of ionospheric effects and
RX drift on LF signals, I recorded about 30 hours worth of the spectrum of
DCF77 at high resolution - see the attachment. The trace is from about 2300
Tuesday until 0800z today, each graduation is 2 hours, and the total
frequency span is about 0.04Hz. The FFT resolution is 0.7 millihertz, and
the frequency stays within +/- 1 FFT bin, ie. within 1 part in 10e8 of the
mean value.
0.0007Hz is 0.004 radians per second, which at 77.5kHz would correspond to
about 2.5m/s, or 9km/hour change in path length. Each FFT requires about 25
minutes worth of data, so if the path length was changing by a few 10s of
km/hour for half an hour or more, one would expect the resultant frequency
shift to be clearly visible. However, there are no clear signs of such a
large shift. The signal is quite "fuzzy" during the hours of darkness,
suggesting to me that small phase changes are occuring over a shorter
period of time. I'm not sure about the glitches in the trace - perhaps
noise of some sort in the system. The receiver AGC is switched on, keeping
the signal level into the sound card virtually constant - the signal level
during the night varied over a range of at least 10dB, showing some sky
wave must have been present.
I can send a bigger image of the plot if anyone would like it.
Cheers, Jim Moritz
73 de M0BMU
|
